Under the hood

How the Clinical Clarity Engine works

Every patient record begins as scattered data across systems, formats, and providers. The xCures Clinical Clarity Engine brings it together, structures it, and turns it into clinical information you can act on.

six steps

From intake to evidence.

Ingest from any source

Medical records can reach the Clinical Clarity Engine through multiple pipelines: direct uploads, system integration, EHRs, HIEs, QHINs, or APIs.

Normalize the formats

Records arrive as C-CDA, FHIR, scanned images, and PDFs. Each format is parsed, run through OCR where needed, and tagged with structured metadata so downstream systems can read it.

Segment and classify

Text is broken into meaningful units using section-based and semantic chunking. Embedding-based classification identifies what each chunk is and where it belongs in the patient’s record.

Retrieve and extract

Targeted semantic search surfaces the right documents. Performance thresholds determine when an extractor graduates from development to validation to large-scale deployment.

Standardize & contextualize

LLM-extracted data is mapped to FHIR R4 resources, normalized to OMOP concepts, and merged with tabular C-CDA data. The result is a comprehensive, standardized longitudinal record.

Verify with evidence

A checklist system interrogates the documents to extract patient-level clinical assertions across the assembled record. Outputs are structured variables accompanied by source-document citations.

Bring clarity to your data.

Walk through xCures with our team – we’ll show how the pipeline works on the data you already have.

Request a demo

Frequently asked.

What data formats does xCures accept?

xCures® accepts records in any format clinical data arrives in: FHIR, C-CDA, scanned PDFs, images, and direct uploads. All source formats are parsed, OCR-processed where needed, and normalized to FHIR R4 for downstream use. Organizations with existing data infrastructure can bring their own data through direct ingestion. No forced migration or single pipeline is required.

How does xCures access patient records?

xCures retrieves patient records through three paths. TEFCA and Carequality networks support treatment-based use cases on behalf of healthcare providers. Individual Access Services (IAS) enables individual-authorized retrieval for digital health and individual-initiated workflows. Direct upload supports organizations bringing their own data. xCures is HITRUST certified and participates in the national interoperability framework.

How long does it take to get a decision-ready output?

For most checklist-based workflows, xCures returns a decision-ready output in minutes. Comprehensive patient history assembly typically completes within the same session. Timelines vary by record complexity, retrieval network availability, and number of care locations. The team can provide workflow-specific estimates during a demo using a synthetic patient record.

Is xCures available across the entire United States?

Yes. xCures operates across all 50 states, reaching more than 95% of patients in the United States through QHINs, HIEs, and direct integrations. The engine processes data from over 550,000 care locations, spanning hospitals, specialty practices, diagnostic labs, and outpatient facilities. National coverage is available today.